How Much of Your Index Fund Is Tied to AI? (2026)

Guides
Moneda Guides: an illustration of the Moneda face

Short answer

  • The count is wide, the weight is not. The S&P 500 holds 503 stocks, but its ten largest made up 37.3% of the index on 30 October 2026. About 2% of the stocks held more than a third of the weight.
  • Much of that top end is tied to AI. By our reading of their own filings, six of the largest companies make AI chips, rent out AI computing or invest heavily to build it. Together they made up about 26% of the index on 3 November 2026.
  • You can check this for most funds. Look past the number of holdings to the top ten weights and the sector split. That gives a clearer picture of what you own.

An S&P 500 index fund (a fund that aims to hold the companies on the index's list, in roughly the proportions the index sets) sounds like the definition of spreading your money out: about 500 large US companies in one go.

The label is accurate about the count. It says much less about the weight. On 30 October 2026, the index's ten largest stocks made up 37.3% of it, according to S&P Dow Jones Indices, which publishes the index. The remaining 493 made up the rest.

This guide explains how a 500-company index ends up leaning on ten names, how many of those names are tied to the same AI spending, and how to see the same picture in a fund you own or are looking at.

Count versus weight: how 500 companies can lean on ten

Many of the best-known indices weight each company by its market value, also called market capitalisation: roughly, the share price multiplied by the number of shares available to investors. The bigger the company, the bigger its slice of the index.

Imagine a small index of ten companies, where one is worth as much as the other nine put together. Put €100 into a fund that tracks it and about €50 goes into that one company, while each of the other nine gets about €5.56. The label says ten companies. Half of your money sits in one.

Scale that up to 500 and the same thing happens, less visibly: a company worth four trillion dollars gets roughly a thousand times the weight of one worth four billion. In the S&P 500 on 30 October 2026, the single largest company was 7.7% of the index. By sector, information technology alone made up 37.4% of the index on the same date, according to S&P Dow Jones Indices. Add communication services, the sector where GICS (the sector classification system developed by S&P and MSCI) places Alphabet and Meta, and it is about 46%.

This is how the method is designed to work. Market-cap weighting means an index broadly mirrors the market as it is, and a fund tracking it needs relatively little trading to keep up, which helps keep costs down. It also means the index becomes more concentrated when the largest companies grow faster than the rest.

Which of the biggest companies are tied to AI?

On 30 October 2026, the ten largest lines in the S&P 500 belonged to nine companies. Alphabet, Google's parent company, appears twice because two classes of its shares are listed. In order of size, the nine were Apple, NVIDIA, Microsoft, Amazon, Alphabet, Broadcom, Meta, Eli Lilly and Tesla.

Here is our summary of how each company's own results and filings describe its link to AI. Figures are as the companies reported them, for the periods shown.

Company Link to AI, from its own results and filings
NVIDIA Makes AI chips. Its data centre business brought in about 92% of its revenue in the quarter to 26 July 2026.
Broadcom Makes chips for AI, including custom chips built for individual customers. Revenue from AI chips was about 56% of its total in the quarter to 2 August 2026.
Microsoft Rents out AI computing through Azure, its cloud business. Its capital spending, including finance leases, was $45.2 billion in the quarter to 30 September 2026.
Amazon Rents out AI computing through AWS, its cloud arm. It spent $77.3 billion on property and equipment in the quarter to 30 September 2026, and says the rise in its spending over the past year mainly reflects investment in AI.
Alphabet Rents out AI computing through Google Cloud, and says it now also offers its own AI chip systems to other companies. Its capital spending was $51.4 billion in the quarter to 30 September 2026.
Meta Invests in AI infrastructure. Its capital spending, including finance leases, was $34.9 billion in the quarter to 30 September 2026.
Apple Says it has invested significantly in AI features and the infrastructure behind them, without breaking out the amount.
Tesla Says it is building its own AI training computers for its self-driving and robotics work, without breaking out the amount.
Eli Lilly A drugmaker. We found no AI link described in its latest quarterly results beyond standard legal wording.

On our grouping, the first six make AI chips, rent out AI computing or invest heavily to build it. Based on index weights published by State Street Global Advisors, they made up about 26% of the S&P 500 between them on 3 November 2026. Add Apple and Tesla, whose filings also describe AI investment, and it is about 36%.

Their businesses are also connected. NVIDIA's figures show that about 55% of its data centre revenue in the quarter to 26 July 2026 came from cloud providers and the largest internet companies, and two customers it does not name accounted for more than a third of its total revenue in its last full financial year. On the other side, the latest guidance from Microsoft, Amazon, Alphabet and Meta adds up to more than $700 billion of capital spending between them in 2026, much of which they attribute to AI computing and the data centres that house it. Guidance can change.

In other words, a sizeable part of the index's weight is linked to the same spending: the companies paying for AI capacity, and the companies supplying it. Several of these companies list among the risks in their annual reports that their AI investment may not earn the return they expect, and both chipmakers note that a large share of their revenue comes from a small number of customers.

Is it the same in other indices?

Broadly, yes, and in the Nasdaq-100 more so.

The Nasdaq-100 follows the 100 largest non-financial companies listed on the Nasdaq exchange. On 30 September 2026, adding up the weights Nasdaq publishes, its ten largest lines made up about 47% of it. Alongside the familiar names, two more chipmakers, Micron Technology and Advanced Micro Devices, sat in its top ten.

MSCI World is often described as a global index. It holds 1,248 stocks across 23 developed countries, and leaves out emerging markets such as China and India. On 30 October 2026, its ten largest lines made up 26.6% of it, and every one of them belonged to the same nine US companies that lead the S&P 500. US companies made up 72.8% of the index overall. A fund tracking it would hold roughly three quarters of its money in US companies, and about a quarter in those same nine names.

When concentration helped, and when it hurt

Concentration is not new, and it has worked in both directions. Estimates differ slightly between sources, depending on the date and on how share classes are counted.

How it has changed. From 1990 to 2015, the ten largest companies in the S&P 500 made up between about 18% and 23% of it, according to RBC Wealth Management, using FactSet data. At the height of the dot-com boom in 2000 the share reached about 27%, Goldman Sachs Research estimated. By mid-2025 it was almost 40%, the highest since the mid-1960s, according to S&P Dow Jones Indices.

When the biggest carried the index. According to S&P Dow Jones Indices, the S&P 500 returned 26.29% in 2023, including dividends, and would have returned 9.94% without the so-called Magnificent Seven (Apple, Microsoft, Alphabet, Amazon, NVIDIA, Meta and Tesla). In 2024 it returned 25.02%, or 11.75% without them, and the seven accounted for 53.1% of the year's return. Owning the whole index meant sharing in those gains.

When the biggest weighed on it. At the end of 1999, the three largest companies in the S&P 500 were Microsoft, General Electric and Cisco. From 2000 to 2013, each of the three had a negative average annual return, while the index as a whole gained 3.6% a year, according to an analysis by Morgan Stanley Investment Management. The index itself fell three years running, including dividends, in 2000, 2001 and 2002, according to data compiled by Aswath Damodaran at NYU Stern. In 2022 the S&P 500's price fell 19.44%. J.P. Morgan Asset Management estimates that the Magnificent Seven fell about 40% as a group that year, against about 12% for the other 493 stocks.

October 2026. TheStreet reported that the tech-heavy Nasdaq Composite fell 5.3% in October, its worst month since March 2025, as chip stocks declined. Over the same month the S&P 500's price fell 0.93%, according to S&P Dow Jones Indices, and MSCI World fell 0.75% in US dollars, including dividends, according to MSCI.

Early November 2026. On 5 November the S&P 500's price fell 4.36% in a single day, according to S&P Dow Jones Indices, as AI and chip stocks led a broad sell-off. The Associated Press reported that the Nasdaq Composite fell about 5.5% the same day, and that it was the S&P 500's worst day since April 2025.

Past performance is not a reliable indicator of future results, and none of this says what comes next. It shows that the way a fund is weighted shapes how it behaves, in both directions.

How to see what your own fund owns

Three checks, usually a minute or two each with a fund's factsheet (the short summary most funds publish):

  1. Look through the fund. A fund is a container. Find its top ten holdings and add up their weights. That number often tells you more about concentration than the count of companies on the cover.
  2. Read the sector weights. Sectors group companies by what they do: information technology, financials, health care and so on. Classification can hide links, though. Under GICS, Alphabet and Meta sit in communication services and Amazon in consumer discretionary, even though all three report large investments in AI.
  3. Work out how many holdings you effectively own. The effective number of holdings turns a list of weights into one figure: how many equal-sized holdings would give the same spread. Ten holdings at 10% each give an effective number of ten. Ten holdings where one takes half and the other nine share the rest equally give about 3.6. It is a quick way to compare a fund's label with its shape.

One fund can be a single line in your portfolio and hundreds of companies inside it. Our guide to researching an ETF goes through a factsheet step by step, and AI portfolio trackers can help bring your holdings into one view.

See inside your funds with Moneda

Moneda puts this research where you invest. Moneda AI does the analysis and shows the sources behind its figures, and you make every decision.

  • Fund pages show what is inside. ETF pages in Moneda show a selection of a fund's largest holdings and, where the data is available, its sector and regional weights, so the look-through is quick.
  • Ask about a fund. Ask Moneda AI what a fund holds, how much of it sits in its ten largest companies, or how those companies describe their links to AI. It works from market data and company filings and shows its sources, so you can check any figure before relying on it.
  • Your portfolio score includes concentration. The score rates a portfolio's return, risk and concentration on a 1 to 7 scale, and measures concentration with the effective number of holdings and the weight of the largest one, so you can see how evenly your money is spread across what you hold. Scores are informational, not investment advice.
  • See how much sits in one company or sector. Moneda AI can show how much of what you hold in Moneda sits in a single company, counting both the shares you own directly and those you hold through your funds, and flags a position once it makes up a large share. For funds it uses the largest holdings each fund publishes, so the figure is a minimum. Your daily briefing also notes when one sector makes up a large part of your portfolio. Money you hold elsewhere is not included.

In your portfolio score a fund counts as one holding, so the fund pages and Moneda AI are where you see inside it. To build a portfolio from the companies you choose, our guide to building a portfolio with AI walks through it, and where available you can invest in US stocks and ETFs, with research built in.

The label counts the companies. The weights tell you what you own.

Frequently asked questions

Does an S&P 500 index fund spread my money evenly?

No. It spreads it in proportion to each company's market value, roughly its share price multiplied by the shares available to investors. The index holds 503 stocks across 11 sectors, but on 30 October 2026 its ten largest made up 37.3% of it, and the single largest was 7.7%.

What counts as tied to AI in this guide?

A company that, in its own results or filings, makes AI chips, rents out AI computing through a cloud business, or invests heavily in AI infrastructure. This is our own grouping, and others may draw the line differently. It describes what the business does. It is not a view on the company's prospects or value.

What is the effective number of holdings?

A single figure for how spread out a portfolio is: the number of equal-sized holdings that would give the same spread. It falls as one or a few holdings grow larger, even if the total number of holdings stays the same.

Sources

Figures are as of the dates shown and will change over time.

This guide is for information and education only and is not investment advice. Companies and indices are named for illustration only, and nothing here is an offer or a solicitation to invest in any of them. Index and company names belong to their owners, and Moneda is not affiliated with or endorsed by them. Figures come from the third-party sources listed, as of the dates shown, and may since have changed. Past performance is not a reliable indicator of future results. The value of investments can fall as well as rise, and you may get back less than you put in. Moneda is a financial technology company, not a bank. Brokerage accounts are provided by licensed partners, and some features offered by regulated third parties on Moneda are only available in selected jurisdictions. Moneda AI acts only within your own account and your own approvals, and it does not provide personalised investment advice.

Share this post